Digital-to-optical converter was developed specially for pile-up and deadtime characterization of nuclear devices, but its application is probably wider. The converter generates 10 ns-wide pulses in visible (660 nm) range, with 8-bit resolution of the optical power, the data rate is up to 20 Ms/s. The output optical pulses are generated by high-luminosity LED diode coupled to the optical fibre. A crucial task of the design is eliminating of influence of the parasitic capacitances in fast switching mode.

The use of the Allan variance for the characterisation of the jitter error in analog-to-digital converters (ADCs) is proposed. In particular, the Allan variance is a sound basis for defining a figure of merit for jitter errors, diagnosing the jitter noise type, and including a jitter block into a previously proposed ADC model. Experimental results highlighting the effectiveness of the Allan variance in the characterisation of the ADC jitter error are discussed.

In analog-to-digital converters (ADCs), dynamic and memory nonlinear effects can contribute simultaneously to the distortion of the digitised signal. These effects can be modelled and compensated effectively via Volterra filters. The paper deals with a Volterra filter for ADC error correction based on an inverse model. An easy-to-implement correction technique, based on a efficient mathematical model of Volterra filter designed to reduce burden in model definition, in filter identification, and in experimental calibration is proposed. Preliminary simulation and experimental results for integrating ADCs highlight the effectiveness of the proposed modelling and correction approach.

The combination of two methods enable to decrease an influence of testing signal distortion for testing of the dynamical quality of middle resolution AD modules in frequency range from several tens of kHz to several MHz is described. In the first step, the distortion of a generated testing signal is suppressed by applying a low distortion LC filter special designed for this purpose. It decreases all spurious components and also even harmonic components of the generated signal, but further odd harmonic component arise due to applying a ferromagnetic material in the used coils. However, in combination with the modified FFT method (spectrum correction test) these undesirable harmonic components can be digitally corrected.

The paper presents a new methodology, based on the Discrete Fourier Transform of the output, for measuring the integral nonlinearity of an ADC. The new method, unlike the usual histogram test, gives the best polynomial approximation (the “smooth part”) of the integral nonlinearity, which is responsible for the spurious harmonics above the ADC noise floor, and can be performed with great accuracy and repeatability with as few as 8,000 samples, irrespective the ADC resolution. This makes this test by far faster than the histogram test for high-resolution devices.